Comments (3)
Hi, thanks for opening the issue. This Github project is about the TF Lite Support library, rather than a place for "support" for TF Lite. In order to proceed here, can you please open a new issue for "Performance" here:
https://github.com/tensorflow/tensorflow/issues/new/choose
and then fill in the template with the relevant details and assign to me? Cheers!
from tflite-support.
Sorry for the late response.
Such performance behavior you described isn't surprising (:-)) because those quant kernels haven't been optimized as much as those corresponding float kernels on Intel CPUs (i.e. x86_64). But if you try it out on mobile phones, you will find that quant inference generally outperforms their float counterpart.
We have been aware of such performance discrepancy, and we have been working on to optimize these quant kernels on x86_64 platforms. I think more recent TFLite builds (say, the tflite nightly) should start to deliver better performance on x86_64 for quant models. @talumbau will be able to shed more light on this.
from tflite-support.
Thanks for the details @multiverse-tf, I would appreciate more info on this too, if @talumbau could give it a go? 🥇
x86 is falling behind 🗡️ 😆
from tflite-support.
Related Issues (20)
- Tflite training failed in Android HOT 2
- TfLiteGpuDelegate Prepare: Output size is less than input size.
- Error getting native address of native library: task_vision_jni HOT 4
- ModuleNotFoundError: No module named 'tflite_micro' HOT 2
- pip install tflite-model-maker not working HOT 9
- python 3.11 aarch64 support HOT 1
- Android NDK Prefabs for Tensorflow Lite
- how to build tflite-support with basel for use in c++
- GPU Delegation error in Andoid (java.lang.IllegalArgumentException: Internal error: Error applying delegate:)
- No categories present in the resulting list after calling classifier.classify(), even though 20 categories are expected to be there.
- Unknown image file format. One of JPEG, PNG, GIF, BMP required.
- tflite-support or tflite_support_nightly are not working on colab HOT 1
- Modify tflite quantization params
- implementation 'org.tensorflow:tensorflow-lite-support:0.1.0'
- Select TensorFlow op(s), included in the given model, is(are) not supported by this interpreter. Make sure you apply/link the Flex delegate before inference. For the Android, it can be resolved by adding "org.tensorflow:tensorflow-lite-select-tf-ops" dependency. See instructions: https://www.tensorflow.org/lite/guide/ops_select
- I want to ask if TFLImage Searcher is supported on iOS and if so, where can I get a sample?
- Getting "Cannot copy to a TensorFlowLite tensor (serving_default_input_1:0) with 63984 bytes from a Java Buffer with 64000 bytes" error while attempting to pass the Input and Output TensorAudio Buffer to TFLite Interpreter for inference HOT 1
- ImportError from image_utils HOT 4
- Could you please provide a aarch64 tflite-support wheel for python 3.10 HOT 2
- YOLOv8 `tensorflow>=2.14.0` ImportError: generic_type: cannot initialize type "StatusCode": an object with that name is already defined HOT 6
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tflite-support.